An acoustically-motivated spatial prior for under-determined reverberant source separation

Ngoc Duong 1 Emmanuel Vincent 1 Rémi Gribonval 1
1 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : We consider the task of under-determined reverberant audio source separation. We model the contribution of each source to all mixture channels in the time-frequency domain as a zero-mean Gaussian random vector with full-rank spatial covariance matrix. We introduce an inverse Wishart prior over the covariance matrices, whose mean is given by the theory of statistical room acoustics and whose variance is learned from training data. We then derive an Expectation-Maximization (EM) algorithm to estimate the model parameters in the Maximum A Posteriori (MAP) sense given prior knowledge about the microphone spacing and the source positions. This algorithm provides a principled solution to the well-known permutation problem and achieves better separation performance than other algorithms exploiting the same prior knowledge.
Type de document :
Communication dans un congrès
Acoustics, Speech and Signal Processing, IEEE Conference on (ICASSP'11), May 2011, Prague, Czech Republic. 2011, 〈10.1109/ICASSP.2011.5946315〉
Liste complète des métadonnées

Littérature citée [7 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00566868
Contributeur : Ngoc Duong <>
Soumis le : dimanche 20 février 2011 - 19:57:38
Dernière modification le : mercredi 16 mai 2018 - 11:23:03
Document(s) archivé(s) le : samedi 21 mai 2011 - 02:38:42

Fichier

icassp2011.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Ngoc Duong, Emmanuel Vincent, Rémi Gribonval. An acoustically-motivated spatial prior for under-determined reverberant source separation. Acoustics, Speech and Signal Processing, IEEE Conference on (ICASSP'11), May 2011, Prague, Czech Republic. 2011, 〈10.1109/ICASSP.2011.5946315〉. 〈inria-00566868〉

Partager

Métriques

Consultations de la notice

623

Téléchargements de fichiers

233